论文部分内容阅读
提出了一种废杂铜铜液温度软测量方法,用于测量宁波某铜业公司改造后的废杂铜冶炼炉还原阶段的铜液温度。该方法首先通过机理分析找出与铜液温度相关的辅助变量,并且研究这些量之间的数学函数关系,简化后续神经网络结构并确定输入层与隐含层初始权重,将黑箱模型转化成灰箱模型。最后利用人工神经网络理论建立铜液温度软测量模型。通过仿真验证表明,该模型能较好地反映出铜液温度与上述变量之间的关系,并且输入层与隐含层初始权重的确定有助于加快神经网络收敛速度,提高模型精度。
A method for soft temperature measurement of copper scrap liquid was proposed to measure the temperature of copper solution in the reduction stage of a scrap copper smelter after the revamping of a copper company in Ningbo. The method first finds out the auxiliary variables related to the temperature of the copper liquid through the mechanism analysis, studies the mathematical function relations among these quantities, simplifies the structure of the subsequent neural network and determines the initial weights of the input layer and the hidden layer, converts the black box model into gray Box model. Finally, artificial neural network theory is used to establish a soft sensor model of copper liquid temperature. The simulation results show that the model can well reflect the relationship between the temperature of copper liquid and the above variables, and the initial weights of input layer and hidden layer help to accelerate the convergence speed of neural network and improve the model accuracy.